SCDb: an integrated database of stomach cancer

1. An integrated SC database, SCDb, was constructed.

2. SC-related genes, miRNAs, and SNPs were identified.

3. KM survival curves of 20,264 genes were generated.

4. Gene_disease, CNV, methylation, drug and TF analyses were performed.

5. Convenient links of the String and GENSCAN databases are provided in the SCDb.

Stomach cancer (SC, also named as gastric cancer) is a type of cancer, which is derived from the stomach mucous membrane [1]. According to the GLOBOCAN 2018 data, SC ranks as the fifth most common neoplasm and the third most leading cause of cancer deaths worldwide, with an estimated count of 783,000 deaths per year [2]. SC is known to reach an advanced stage with relatively poor prognosis due to the non-specific symptoms or no noticeable symptoms appearing in the early stages [3]. The early symptoms of SC include upper abdominal pain, heartburn, loss of appetite and nausea, and the later symptoms include yellowing of the skin and whites of the eyes, weight loss, difficulty in swallowing and excessive vomiting [4]. Besides this, SC also show metastasis from stomach to other tissues or organs, especially the lungs, liver, lining of the abdomen, bones, and lymph nodes [5]. In most of the SC cases (more than 60%), it has been shown to be induced by Helicobacter pylori infection [6‐8], whereas other causes include smoking, eating pickled vegetables and genetic syndromes [7]. SC is difficult to cure because the patients that are diagnosed with the disease usually have reached an advanced stage [9]. The conventional treatments for SC include surgery [10], radiation therapy, and/or chemotherapy [11].

Although many researchers have performed a series of genomics, proteomics, transcriptomics, and epidemiological studies with regard to SC [12‐15], there is only one available database of human gastric cancer, which is the Database of Human Gastric Cancer (DBGC, http://​bminfor.​tongji.​edu.​cn/​dbgc/​index.​do) [16]. The DBGC database has integrated human gastric cancer-related biomarkers, drug-sensitive genes, mutations, transcriptomics projects and proteomics projects from different sources, however, some useful information is still excluded from it, as the datasets are greatly dispersive and heterogeneous [16]. Besides this, there is another database Online Mendelian Inheritance in Man (OMIM, http://​www.​ncbi.​nlm.​nih.​gov/​omim) [17], which is an authoritative, comprehensive and timely database that involves the relationship between genotype and phenotype of all human genetic disorders. The miR2Disease [18] and HMDD [19] databases contain comprehensive information about the miRNAs that are related to multiple human diseases. ClinVar database (http://​www.​ncbi.​nlm.​nih.​gov/​clinvar/​) provides a repository of relationships among important variants and phenotypes in medical [20]. The above databases majorly focus on molecular mechanisms of various diseases, and not just on SC. Therefore, it is of great importance to develop an integrated SC-specific database which will include gene, gene-disease, miRNA, miRNA_disease, copy number variations (CNVs), single nucleotide polymorphism (SNP), SNP-disease, methylation, drug and transcription factors (TFs).

In this study, we constructed an integrated database of stomach cancer, SCDb, (available at http://​www.​stomachcancerdb.​org/​) by retrieving the databases and literature mining and by performing bioinformatics analysis of the publicly available datasets. This human SC database might help researchers to investigate and provide more information about the human SC-related molecules from several clinical aspects.

As there are non-specific symptoms or no such noticeable symptoms observed in early stages of SC, newly diagnosed SC cases usually reach an advanced stage and are thus difficult to cure. To better understand the pathogenesis of SC, we developed the SCDb database that includes information on SC-related genes, gene_disease, miRNA, miRNA_disease, CNV, SNP, SNP_disease, methylation, drug, TF and KM survival curves. All this information was retrieved by analyzing the microarray datasets and by mining the literature. Information on SC-related genes (eg. gene symbol, up−/down- regulation, GEO ID and PUBM ID), gene_disease (eg. gene symbol and gene_disease), miRNA (eg. gene symbol and miRNA symbol), miRNA_disease (eg. gene symbol and miRNA symbol), CNV (eg. gene symbol and CNV), SNP (eg. gene symbol and SNP), SNP_disease (eg. gene symbol and SNP_disease), methylation (eg. gene symbol and methylation), drug (eg. gene symbol and drug), and TF (eg. gene symbol and TF) were integrated into this database. At present, the database includes information of 9990 SC-related genes, 65 confirmed SC-related genes, 457 miRNAs, 20 SC-related miRNAs, 1570 SNPs, 108 SC-related SNPs, 419 TFs, 44,605 CNVs, 3404 drug-associated genes, 63 genes with methylation and KM survival curves of 20,264 genes.

Compared to the previously established DBGC database [16], the SCDb database has several advantages: (1) SCDb database includes not just previously established information i.e. specifically, by performing the analyses using the microarray datasets, and RNA-seq datasets, novel genes, miRNAs, and SNPs were identified, which can further contribute to the determination of new directions for SC-related research; (2) SCDb database provides detailed regulatory information, for instance, possible TF-gene and miRNA-gene pairs associated with SC might also be identified based on the SC-related genes information; (3) a comprehensive analysis was performed for the SC-related genes, and various other data were integrated into the database, including the information on gene, gene_disease, miRNA, miRNA_disease, CNV, SNP, SNP_disease, methylation, drug, TF, and KM survival curves; (4) SCDb provides a search engine for tools, including String and GENSCAN, and thus protein-protein interaction analysis and gene prediction for unknown sequences can also be performed using SCDb; (5) SCDb provides search engines for Query, Browse, and Summary. Therefore, we can not only perform a search for gene, gene_disease, miRNA, miRNA_disease, CNV, SNP, SNP_disease, methylation, drug, TF terms, and KM survival curves in detail but can also obtain all the corresponding information of each term that is included in the SCDb database and all information related to one gene.

However, the gene expression data that were collected from multiple publicly available microarray datasets and more details of these datasets, such as number of patients, ethnicity of patients, and how the samples were prepared were not provided, which might be potential limiting factors influencing our results. Moreover, with the advancements in sequencing techniques, the microarray data about SC might not be constantly updated in GEO, and therefore, next-generation data about SC should be obtained, which might provide new insights into SC biology, and should be added if available. Furthermore, this established SCDb database does not provide any information on gene expression based on clinical parameters, such as age, gender, histological or molecular subtypes, tumor stage or grading, and prior therapies. Lastly, we did not conduct the analysis of the correlation between cancer progression stages with gene expression data as well as the multivariate analysis to detect more specific prognostic markers for survival. Considering these limitations, we plan to update the database periodically to continuously improve the quality of the SC-related data and the corresponding functions, thus keeping a track of improvements and advancements in this field.

In conclusion, the SCDb database provides a comprehensive resource for performing research on human SC. We believe that SCDb will be a helpful database for biologists and pharmacologists in the field of SC research, and will promote the studies to better understand the molecular mechanisms of this disease.